# Multi-modality **URL:** https://discuss.vllm.ai/c/features/multi-modality/15.md [Latest](https://discuss.vllm.ai/latest.md) · [Categories](https://discuss.vllm.ai/categories.md) --- ## [About the Multi-modality category](https://discuss.vllm.ai/t/about-the-multi-modality-category/83)
**Author:** [@DarkLight1337](https://discuss.vllm.ai/u/DarkLight1337)\ **Replies:** 0
--- ## [Compressed Multimodal embeddings inputs](https://discuss.vllm.ai/t/compressed-multimodal-embeddings-inputs/2469)
**Author:** [@guillaume.guy](https://discuss.vllm.ai/u/guillaume.guy)\ **Replies:** 1\ **Last updated:** [March 18, 2026, 5:49pm UTC](https://discuss.vllm.ai/t/compressed-multimodal-embeddings-inputs/2469 "2026-03-18T17:49:24Z")
vLLM supports multimodal inputs through the multi\_modal\_data argument. This works well but the argument tends to be quite large. For QWEN3VL-4B with a photo of 384x384, we are talking about 2MB payload (500k floats). On… --- ## [How do I precompute multimodal embeddings?](https://discuss.vllm.ai/t/how-do-i-precompute-multimodal-embeddings/2294)
**Author:** [@kevinpw](https://discuss.vllm.ai/u/kevinpw)\ **Replies:** 5\ **Last updated:** [February 2, 2026, 12:11pm UTC](https://discuss.vllm.ai/t/how-do-i-precompute-multimodal-embeddings/2294 "2026-02-02T12:11:18Z")
I want to efficiently generate embeddings with Qwen3-VL-Embedding-8B for a single image using multiple different text instructions. Embedding generation is relatively slow because the model expects instructions in the sy… --- ## [Is there a plan for EVS to support Qwen3VL in response to the issue of sparse video tokens?](https://discuss.vllm.ai/t/is-there-a-plan-for-evs-to-support-qwen3vl-in-response-to-the-issue-of-sparse-video-tokens/2215)
**Author:** [@XSH](https://discuss.vllm.ai/u/XSH)\ **Replies:** 1\ **Last updated:** [January 13, 2026, 9:31pm UTC](https://discuss.vllm.ai/t/is-there-a-plan-for-evs-to-support-qwen3vl-in-response-to-the-issue-of-sparse-video-tokens/2215 "2026-01-13T21:31:35Z")
Currently, EVS supports QWen2.5vl. Are there plans to support QWen3vl? When trying to upgrade EVS to QWen3VL, the get\_video\_replacement\_qwen3vl method in the qwen3\_vl.py file first calculates the number of video\_token\_ids… --- ## [Qwen2.5 VL开启flashinfer失败](https://discuss.vllm.ai/t/qwen2-5-vl-flashinfer/1999)
**Author:** [@sxcnmslll](https://discuss.vllm.ai/u/sxcnmslll)\ **Replies:** 5\ **Last updated:** [November 24, 2025, 8:15am UTC](https://discuss.vllm.ai/t/qwen2-5-vl-flashinfer/1999 "2025-11-24T08:15:51Z")
Qwen2.5 VL开启flashinfer失败，指令如下： VLLM\_ATTENTION\_BACKEND=FLASHINFER \\ TMPDIR=/msswift\_tmp/tmp \\ vllm serve Qwen/Qwen2.5-VL-3B-Instruct \\ --data-parallel-size 8 \\ --gpu-memory-utilization 0.95 \\ --port 9014 \\ --no-e… --- ## [Custom modality](https://discuss.vllm.ai/t/custom-modality/1916)
**Author:** [@Mateusz-Smolarczyk-A](https://discuss.vllm.ai/u/Mateusz-Smolarczyk-A)\ **Replies:** 3\ **Last updated:** [November 14, 2025, 1:41pm UTC](https://discuss.vllm.ai/t/custom-modality/1916 "2025-11-14T13:41:17Z")
Hello, I’m new to vLLM. I’m currently working on extending a multimodal model within vLLM, and I would like to add support for a custom modality (for example ECG (electrocardiogram) signals). The documentation describe… --- ## [Prompt\_embeds usage in vllm openai completion api](https://discuss.vllm.ai/t/prompt-embeds-usage-in-vllm-openai-completion-api/825)
**Author:** [@ijs](https://discuss.vllm.ai/u/ijs)\ **Replies:** 4\ **Last updated:** [June 17, 2025, 1:15pm UTC](https://discuss.vllm.ai/t/prompt-embeds-usage-in-vllm-openai-completion-api/825 "2025-06-17T13:15:41Z")
Hi, I’m trying to validate support for embeds usage as prompts. To verify this I’ve been utilizing a small model like facebook/opt-125m, giving it a textual prompt, mapping the prompt to corresponding embeddings, and ve…